Second line drug therapy for rheumatoid arthritis (RA) has evolved based on the drugs introduced over time and on clinical judgement. There has been little scientific evidence of these drugs' relative efficacy or toxicity in part because most comparative trials have been too small to detect differences. Furthermore, while it has been suggested that certain subgroups of patients (e.g. those with recent onset disease) respond better to second line drugs than others, there has been little evidence supporting subgroup differences in response. The general objective of this study is to use raw data and published data from clinical trials to evaluate the efficacy and toxicity of treatments for RA. There are two components of this project, a raw data archive in which data is patient specific and a published trial metaanalysis in which the treatment group is studied. Both components build on our prior work. First, we will use our developing data of raw data from RA trials to ask a series of questions about the response of individual patients to second line drugs. This includes two specific aims: first, we will investigate drug efficacy and toxicity in subsets of patients including those with mild disease and those with recent onset disease; second, we will perform intent to treat all analyses to determine the extent to which all patients entering into trials improved. The second component of this project involves an expansion of our metaanalysis of published clinical trials in RA. We will pursue three specific aims: first, we will evaluate the efficacy, toxicity and therapeutic indexes for second line drugs we have not previously studied, including combination therapy and cyclosporine; second, we will gather additional information about previously studied drugs including methotrexate and sulfasalazine; third, we will compare the traditional metaanalytic method with our novel approach. The results of this project will provide new information on the relative value of second line drugs in RA. With regular updates, we will test newly introduced drugs to see how they compare with traditional ones and with the data archive, we will evaluate whether different types of patients respond differently to these drugs.